arXiv subjects

Yang Chi

Publications and source records attributed to Yang Chi.

5 recordsLinked to original sources

arXiv2026

Laguna M.1/XS.2 Technical Report

We present Laguna M.1 and Laguna XS.2, two Mixture-of-Experts foundation models built for long-horizon, agentic coding: M.1 has $225.8$B total parameters ($23.4$B activated per token) and XS.2 has $33.4$B total ($3$B activated). Both models were trained from scratch end-to-end inside the same internal system that we refer to as our Model Factory: a tightly-integrated stack of versioned data, training, evaluation, and inference components that turn model development into an industrial process. We describe the principles and design choices of the Model Factory and also detail the end-to-end training process of our models, throughout pre-training data and architecture, post-training stages, evaluation, and quantization. On agentic software engineering and terminal benchmarks (SWE-bench Verified, SWE-bench Multilingual, SWE-Bench Pro, and Terminal-Bench 2.0) M.1 and XS.2 are competitive with state-of-the-art open models in their respective weight classes. Laguna XS.2 weights are released under Apache~2.0 at https://huggingface.co/collections/poolside/laguna-xs2.

cs.AI↗

arXiv2024

Evaluation of Microscopic Origins in Nonlinear Optical Crystals: Based on Rigorous Atomic Space Tessellating

Clarifying the contribution of various atoms and structural units formed in nonlinear optical (NLO) crystal materials to macroscopic optical response is crucial for NLO crystal design. In this work, the rigorous atomic space tessellating (AST) method without any empirical parameters is introduced to evaluate the correlated contributions in wavelength dependent optical responses. The NLO crystal CsSbF2SO4 with significant asymmetric electron distribution of Sb3+ was selected as the case study. This study indicates that there are significant differences in atomic and dipole moment contributions between static and dynamic second harmonic generation (SHG) susceptibility. Using only the information obtained from NLO-active unit analysis of static SHG susceptibilities to explain the NLO phenomenon in actual photoexcitation can even lead to qualitative errors. Among all the transitions in SHG effects, the on-site transition of Sb has the largest component, while the O-Sb off-site transition with multiple transition channels contributes the most to the macroscopic SHG susceptibilities. The relevant methods have also been extended to linear optics and used to study first-order susceptibilities anisotropy, which is important for achieving phase matching. Sb3+ contributes significantly to the linear optical anisotropy with the same sign as SO42-.

physics.optics↗

arXiv2023

Representing Interlingual Meaning in Lexical Databases

In today's multilingual lexical databases, the majority of the world's languages are under-represented. Beyond a mere issue of resource incompleteness, we show that existing lexical databases have structural limitations that result in a reduced expressivity on culturally-specific words and in mapping them across languages. In particular, the lexical meaning space of dominant languages, such as English, is represented more accurately while linguistically or culturally diverse languages are mapped in an approximate manner. Our paper assesses state-of-the-art multilingual lexical databases and evaluates their strengths and limitations with respect to their expressivity on lexical phenomena of linguistic diversity.

cs.CL↗

arXiv2022

CharFormer: A Glyph Fusion based Attentive Framework for High-precision Character Image Denoising

Degraded images commonly exist in the general sources of character images, leading to unsatisfactory character recognition results. Existing methods have dedicated efforts to restoring degraded character images. However, the denoising results obtained by these methods do not appear to improve character recognition performance. This is mainly because current methods only focus on pixel-level information and ignore critical features of a character, such as its glyph, resulting in character-glyph damage during the denoising process. In this paper, we introduce a novel generic framework based on glyph fusion and attention mechanisms, i.e., CharFormer, for precisely recovering character images without changing their inherent glyphs. Unlike existing frameworks, CharFormer introduces a parallel target task for capturing additional information and injecting it into the image denoising backbone, which will maintain the consistency of character glyphs during character image denoising. Moreover, we utilize attention-based networks for global-local feature interaction, which will help to deal with blind denoising and enhance denoising performance. We compare CharFormer with state-of-the-art methods on multiple datasets. The experimental results show the superiority of CharFormer quantitatively and qualitatively.

cs.CV↗

arXiv2014

TCP-Forward: Fast and Reliable TCP Variant for Wireless Networks

The congestion control algorithms in TCP may incur inferior performance in a lossy network context like wireless networks. Previous works have shown that random linear network coding can improve the throughput of TCP in such networks, although it introduces extra decoding delay at the destination. In this paper we try to alleviate the decoding delay by replacing random linear network coding with LT Codes. Due to the inherent difference between linear network coding and Fountain Codes, such replacement is not as simple as it sounds. We conquer some practical problems and come up with TCP-Forward, a new TCP variant which offers many properties that TCP as a streaming transport protocol should offer. Our performance evaluation shows TCP-Forward provides better performance than previous works.

cs.NI↗